By Martin Brampton, 27 May 2003 08:12
COMMENT Hardware stalwarts such as IBM, Sun and most recently - and most vocally - Intel have been touting 64-bit computing for a while. But while it's a natural evolution, Martin Brampton asks, isn't it also a non-issue for most users? Intel has been talking up 64-bit computing. Ironically, while the processor giant has struggled to attract any interest, AMDs 64-bit launch looks like making big waves in a number of directions. Perversely, the supposedly modern PC world is largely a prisoner of its own history. The original IBM PC was built round an Intel processor that was never designed for general purpose computing but meant as an embedded chip in peripheral devices. The momentum quickly gained by the PC was based on binary compatibility of software, a fact that tied everyone to a given instruction set. It was only with the 80386 that Intel created a processor in the knowledge that personal computers were a significant force. While the design was a major step forward, it was several years before operating systems and applications took full advantage of its capabilities. And most of the instruction set was still tied to earlier devices. Over the years, there have been various attempts to break the historical chains. The PowerPC was touted as an alternative processor and Windows NT was created with a software layer that insulated it from hardware dependencies. However, the insulation applied only to the operating system and all popular software continued to be shipped as binary code, locked to a specific hardware instruction set. Faced with stagnant sales, Microsoft soon dropped support for non-Intel processors. Now the shift to 64-bit processing is hardly as urgent as was the shift to 32-bit enabled by the 80386. Both operating systems and applications were rendered simpler by the extended addressing that became easily available with 32-bits. Despite Intels huffing and puffing, the only real need for 64-bit processing is largely confined to very large databases and high-end graphics processing. In this context, very large databases are truly enormous. While there are important examples, much of the vast quantity of stored data is quite comfortable with 32-bit processing. The inevitable result has been that sales of Intel 64-bit processors have been pretty dull. Intel has been open about the fact that it will be some years before its 64-bit devices will catch up with the speed of 32-bit processors for the general run of software. AMD has caused a splash in this otherwise turgid pond by announcing a processor that will provide both 64-bit and 32-bit processing at top speed. Buyers looking for first class performance and a degree of future proofing will be very tempted by this approach. Intel is rumoured to be sufficiently impressed to have started planning a similar processor to protect its own market share. Dramatic effects in the software market could result from a sudden awakening of the 64-bit market. Microsoft has moved relatively slowly, reflecting the sluggish market. Linux has long been poised to take advantage of the new technology but nothing much has happened. Now we can expect an interesting battle as the Linux proponents aim to push the operating system into a wider range of server markets, beyond the traditional internet-related sectors. On the downside, it is hard not to be concerned about what we might call software drag. As hardware has become spectacularly more powerful and memory plentiful and cheap, software seems to have absorbed much of the gain. Certainly, there have been advances in functionality but it is not at all obvious that they match the huge gains offered by hardware technology. From the buyers perspective, the worst outcome would be that the ready availability of fast 64-bit processors will combine with continuing falls in the price of memory to support a new generation of ever more bloated software. Currently, there are signs of corporate resistance to frequent hardware renewal. It will be a wasted opportunity if 64-bit processing turns out to be nothing more than a new twist to the upgrade spiral. ** Martin Brampton is a director and founder of Black Sheep Research (www.black-sheep-research.co.uk ), an independent consultancy providing research, writing and speaking services on a wide range of business and technology subjects. Martin was previously a director at Bloor Research, and has worked with IT as a user and analyst for over 20 years. He can be contacted at silicon@black-sheep-research.co.uk. For past Devil's Advocate columns see the links below, or type 'Devil' into our search engine.

In order to post a comment you need to be registered and logged in.
Log in or create your silicon.com account below